On the possibility of yet a third kinetochore system in the protist phylum Euglenozoa

Abstract

Transmission of genetic material from one generation to the next is a fundamental feature of all living cells. In eukaryotes, a macromolecular complex called the kinetochore plays crucial roles during chromosome segregation by linking chromosomes to spindle microtubules. Little is known about this process in evolutionarily diverse protists. Within the supergroup Discoba, Euglenozoa forms a speciose group of unicellular flagellates-kinetoplastids, euglenids, and diplonemids. Kinetoplastids have an unconventional kinetochore system, while euglenids have subunits that are conserved among most eukaryotes. For diplonemids, a group of extremely diverse and abundant marine flagellates, it remains unclear what kind of kinetochores are present. Here, we employed deep homology detection protocols using profile-versus-profile Hidden Markov Model searches and AlphaFold-based structural comparisons to detect homologies that might have been previously missed. Interestingly, we still could not detect orthologs for most of the kinetoplastid or canonical kinetochore subunits with few exceptions including a putative centromere-specific histone H3 variant (cenH3/CENP-A), the spindle checkpoint protein Mad2, the chromosomal passenger complex members Aurora and INCENP, and broadly conserved proteins like CLK kinase and the meiotic synaptonemal complex proteins SYCP2/3 that also function at kinetoplastid kinetochores. We examined the localization of five candidate kinetochore-associated proteins in the model diplonemid, Paradiplonema papillatum. PpCENP-A shows discrete dots in the nucleus, implying that it is likely a kinetochore component. PpMad2, PpCLK^KKT10/19, PpSYCP2L1^KKT17/18, and PpINCENP reside in the nucleus, but no clear kinetochore localization was observed. Altogether, these results point to the possibility that diplonemids evolved a hitherto unknown type of kinetochore system.

Importance: A macromolecular assembly called the kinetochore is essential for the segregation of genetic material during eukaryotic cell division. Therefore, characterization of kinetochores across species is essential for understanding the mechanisms involved in this key process across the eukaryotic tree of life. In particular, little is known about kinetochores in divergent protists such as Euglenozoa, a group of unicellular flagellates that includes kinetoplastids, euglenids, and diplonemids, the latter being a highly diverse and abundant component of marine plankton. While kinetoplastids have an unconventional kinetochore system and euglenids have a canonical one similar to traditional model eukaryotes, preliminary searches detected neither unconventional nor canonical kinetochore components in diplonemids. Here, we employed state-of-the-art deep homology detection protocols but still could not detect orthologs for the bulk of kinetoplastid-specific nor canonical kinetochore proteins in diplonemids except for a putative centromere-specific histone H3 variant. Our results suggest that diplonemids evolved kinetochores that do not resemble previously known ones.

Keywords: Diplonemea; Kinetoplastea; Paradiplonema; cell division; cenH3/CENP-A; kinetochore.

Fig 1

Kinetochore composition in Euglenozoa. (A) Cartoon of phyletic relationships of the superphylum Discoba with emphasis on the phylum Euglenozoa harboring kinetoplastids, euglenids, and diplonemids. (B) Cartoon of 61 components of the conventional KT (cKT), SAC, and the anaphase promoting complex/cyclosome (APC/C) as inferred to have been present in the ancestral eukaryotes (7). (C) Cartoon of 41 components and structure of the kinetoplastid kinetochore (KKT) in Trypanosoma brucei, also including the SAC and APC/C components. (D) Overview of the twofold strategy taken toward highly sensitive identification of cKT/KKT orthologs in the Paradiplonema papillatum proteome. Schematic overviews of the reciprocal structure-based (left) and sequence-based homology searching strategies (right). In the center, an example of a network graph is shown that is produced based on these comprehensive searches, with green octagonal nodes indicating a cKT/KKT model, a rectangular yellow node, a P. papillatum accession or diplonemid orthologous group (OG), and in circular blue nodes, any non-KT protein from the complete foldomes of any of the representative species (Homo sapiens, Saccharomyces cerevisiae, Arabidopsis thaliana, Trypanosoma brucei). Arrows (edges) between nodes represent homology links colored according to hierarchy in the search results based on E-value (red: best, blue: second-best, dark gray: third-best, and light gray: other hit).

Fig 2

Detection of bona fide kinetochore proteins in P. papillatum. Network graphs showing similarity links (E-value) as inferred by Foldseek or HHsearch, displayed as edges between homologous folds depicted as nodes (left) and structural alignments as produced by Foldseek where available (right). Network graphs represent an individual MCL. Structural alignments show the predicted structure of the relevant P. papillatum protein placed in the cluster to the AF2 structure that was best hit in the Foldseek against the foldomes of representative species, colored in tan. Predicted protein structures are colored by pLDDT score following the standard AF2 palette. Panels represent different functional/structural aspects of the cKT and KKT, as indicated. *MCL clusters obtained from a network based on HHsearch homology inferences, no structural alignment is available for these models as the folds do not contain a clear structural domain. The best-predicted (highest pLDDT score) AF2 model of the P. papillatum accession is shown on the right.

Fig 3

Closest homologs to NDC80/NUF2 and KKT4 are not 1-to-1 orthologs but outparalogs CCDC93/Cluap1 and NUP92, respectively. Network graphs showing homology hits as inferred by Foldseek, displayed as edges between homologous folds depicted as nodes. (A) All nodes connected to NUF2 and NDC80 nodes through either incoming or outgoing edges, and all nodes connected to these nodes with either directionality to the edge (i.e., all nodes until one degree of separation to a NUF2 and/or NDC80 node). Nodes are positioned according to the E-value represented by connecting edges, where shorter edges equate to lower E-value. On the left, a zoom-in of the network around the node of the P. papillatum accession that is the closest homolog by one of the NUF2/NDC80 query models (marked by a red star) and its closely related hits. Schematic representations of the path connecting NDC80 (top) and NUF2 (bottom) query nodes to their respective top P. papillatum accession hit and its subsequent best hits are shown. (B) Similar to panel A, but with nodes connected to KKT4 through maximally one degree of separation.

Fig 4

Multiple kinetochore systems in the supergroup Discoba. Presence/absence pie chart (Coulson plot) matrix of both canonical as well as kinetoplastid-specific kinetochore proteins in the eukaryotic supergroup Discoba, including a projection of relevant evolutionary events for cKT and KKT proteins onto the tree (cartoon left). Kinetoplastea have kinetochores consisting of KKT/KKIP subunits. Euglenida have kinetochores with the conserved Ndc80 complex, CENP-A, the SKA complex, and most likely an active spindle checkpoint (Bub1/BubR1, Mad2, Mad1, Bub3). Diplonemea harbor a CENP-A-like protein, only minimal numbers of KKT/KKIPs (CLK^KKT10/19, SYCP3^KKT16, SYCP2^KKT17/18, KKIP7, and KKIP10) and cKT proteins (p31^COMET, CENP-S/X, Mad1, Mad2). Overall, Euglenozoa appear to have lost and replaced the cKT in multiple steps in the different clades—see before. Ancestrally, Discoba would have had a complex kinetochore system with a full CCAN (Jakobida), Mis12-C, Ndc80-C, DAM-C, and SKA-C, including an active spindle checkpoint and APC/C. Top left: cartoon of conserved cKT/KKT-related proteins present in P. papillatum as projected onto the LECA cKT composition; gray indicates absent, colored proteins indicate present. Top right: Coulson-style plots of cKT/KKT complexes—below: colored parts of the pie chart are present, light gray means these proteins are absent. Numbers indicate the number of paralogs present. Colors correspond to the cartoons of the cKT and KKT shown in Fig. 1. Per major taxonomic groups among Discoba, a representative species is highlighted, including the presence of a cartoon of its cellular outline. Bottom: presence/absence of three model eukaryotes for which kinetochores have been determined: H. sapiens, S. cerevisiae, and A. thaliana.

Fig 5

Endogenous tagging of putative mitotic proteins in P. papillatum. (A) Immunoblot analysis of P. papillatum cell lines expressing a PrA fusion of CENP-A, INCENP, Mad2, CLK^KKT10/19, and SYCP2L1^KKT17/18. Anti-alpha-tubulin was used as a loading control. (B) Western blot analysis of P. papillatum parental cell line and transgenic cell lines expressing CENP-A-V5.

Fig 6

DIPPA_32769 is a putative CENP-A candidate in P. papillatum that forms distinct foci in the nucleus reminiscent of centromeric staining. (A) Multiple sequence alignment of the CENP-A candidate in P. papillatum with histone H3 and CENP-A from various eukaryotes. Top: secondary structure of histone H3-like proteins. Note that DIPPA_32769 has a longer loop 1 (highlighted in the red box) as well as replacement of Gln (arrow), which are characteristic features of CENP-A. Note that DIPPA_03805 has a longer loop 1 but has Gln. CENP-A candidates are present in other diplonemids (Rhynchopus humris and Hemistasia phaeocysticola), and putatively in Prokinetoplastida (Apiculatamorpha spiralis). (B) Immunostaining of CENP-A-PrA visualized by standard fluorescence microscopy. (C) Immunostaining of CENP-A-V5 visualized using a confocal microscope confirming its distinct distribution pattern in the nucleus. Single slices are shown. Image files that have full z-sections are available in File S3, at Figshare. Scale bars: 10 µm.

Fig 7

P. papillatum Mad2 and INCENP are nuclear proteins. (A) Mad2-PrA localizes in the nucleus. Scale bar: 10 µm. Cdc20 in diplonemids (P. papillatum, R. humris, and H. phaeocysticola), but not in kinetoplastids (T. brucei, Trypanosoma cruzi, and Neobodo designis), has a putative Mad2-binding motif. AlphaFold3 prediction of PpCdc20-A (DIPPA_04340) interaction with PpMad2 (DIPPA_17925) via the predicted MIM (62), with an iPTM score of 0.66. The predicted Mad2-Cdc20 interaction indicates that a form of SAC-like signaling is active in Diplonemea. (B) PpINCENP lacks the C-terminal cross-phosphorylation Threonine-Serine-Serine (TSS) site needed for proper activation for Aurora kinases in some model organisms, e.g., humans. Yeasts and Arabidopsis appeared to have lost this motif. AF3 prediction of INCENP interaction with Aurora paralogs pinpointed DIPPA_24328 as the most likely interaction partner (interface Predicted Template Modelling (iPTM) score:0.66). See Fig. S9 for AF3 interaction predictions between INCENP and Aurora kinases paralogs.

Fig 8

SYCP2L1^KKT17/18 and CLK^KKT10/19 are localized to the nucleus. (A) SYCP2L1^KKT17/18 (DIPPA_35871)-PrA localizes in the nucleus. Scale bar: 10 µm. Two paralogs of SYCP2 are present in P. papillatum: DIPPA_35871 (SYCP2L1) and DIPPA_35886 (SYCP2L2). Right: AF3-based predictions of putative closure motifs SYCP2 orthologs that interact with the HORMA protein Hop1, showing specifically the interactions predicted for SYCP2L1 and SYCP2L2. No closure motif was found in SYCP2L1, hinting at a role for SYCP2L1 during mitosis and SYCP2L2 during meiosis (iPTM:0.31). Top right: alignment of verified and predicted closure motifs. (B) CLK^KKT10/19 (DIPPA_05595)-PrA localizes in the nucleus. Scale bar: 10 µm. Phylogenetic tree of the CLK kinase family including both KKT10/19 and the closest paralog in P. papillatum (DIPPA_05595).